Module:scripts: Difference between revisions

From Linguifex
Jump to navigation Jump to search
No edit summary
No edit summary
Tag: Reverted
Line 1: Line 1:
local export = {}
local export = {}
local Script = {}
local Script = {}


--[==[Returns the script code of the language. Example: {{code|lua|"Cyrl"}} for Cyrillic.]==]
function Script:getCode()
function Script:getCode()
return self._code
return self._code
end
end


--[==[Returns the canonical name of the script. This is the name used to represent that script on Wiktionary. Example: {{code|lua|"Cyrillic"}} for Cyrillic.]==]
function Script:getCanonicalName()
function Script:getCanonicalName()
return self._rawData.canonicalName
return self._rawData[1] or self._rawData.canonicalName
end
end


--[==[Returns the display form of the script. For scripts, this is the same as the value returned by <code>:getCategoryName("nocap")</code>, i.e. it reads "NAME script" (e.g. {{code|lua|"Arabic script"}}). For regular and etymology languages, this is the same as the canonical name, and for families, it reads "NAME languages" (e.g. {{code|lua|"Indo-Iranian languages"}}). The displayed text used in <code>:makeCategoryLink</code> is always the same as the display form.]==]
function Script:getDisplayForm()
function Script:getDisplayForm()
return self:getCategoryName("nocap")
return self:getCategoryName("nocap")
end
--[==[Returns the text direction, if any. Currently, left-to-right scripts are unmarked, while most right-to-left scripts have direction specified as {{code|lua|"rtl"}} and Mongolian as {{code|lua|"down"}}.]==]
function Script:getDirection()
return self._rawData.direction
end
end


Line 31: Line 30:
end
end


--[==[Returns the parent of the script. Example: {{code|lua|"Latn"}} for {{code|lua|"Latnx"}} and {{code|lua|"Arab"}} for {{code|lua|"fa-Arab"}}. It returns {{code|lua|"top"}} for scripts without a parent, like {{code|lua|"Latn"}}, {{code|lua|"Grek"}}, etc.]==]
function Script:getParent()
function Script:getParent()
return self._rawData.parent
return self._rawData.parent
end
function Script:getSystemCodes()
if not self._systemCodes then
if type(self._rawData[2]) == "table" then
self._systemCodes = self._rawData[2]
elseif type(self._rawData[2]) == "string" then
self._systemCodes = mw.text.split(self._rawData[2], "%s*,%s*")
else
self._systemCodes = {}
end
end
return self._systemCodes
end
end


Line 40: Line 53:
self._systemObjects = {}
self._systemObjects = {}
for _, sys in ipairs(self._rawData.systems or {}) do
for _, ws in ipairs(self:getSystemCodes()) do
table.insert(self._systemObjects, m_systems.getByCode(sys))
table.insert(self._systemObjects, m_systems.getByCode(ws))
end
end
end
end
Line 52: Line 65:
--end
--end


 
--[==[Given a list of types as strings, returns true if the script has all of them. Possible types are explained in [[Module:scripts/data]].]==]
function Script:getType()
function Script:hasType(...)
return "script"
if not self._type then
self._type = {script = true}
if self._rawData.type then
for _, type in ipairs(mw.text.split(self._rawData.type, "%s*,%s*")) do
self._type[type] = true
end
end
end
for _, type in ipairs{...} do
if not self._type[type] then
return false
end
end
return true
end
end


 
--[==[Returns the name of the main category of that script. Example: {{code|lua|"Cyrillic script"}} for Cyrillic, whose category is at [[:Category:Cyrillic script]].
Unless optional argument <code>nocap</code> is given, the script name at the beginning of the returned value will be capitalized. This capitalization is correct for category names, but not if the script name is lowercase and the returned value of this function is used in the middle of a sentence. (For example, the script with the code <code>Semap</code> has the name <code>"flag semaphore"</code>, which should remain lowercase when used as part of the category name [[:Category:Translingual letters in flag semaphore]] but should be capitalized in [[:Category:Flag semaphore templates]].) If you are considering using <code>getCategoryName("nocap")</code>, use <code>getDisplayForm()</code> instead.]==]
function Script:getCategoryName(nocap)
function Script:getCategoryName(nocap)
local name = self._rawData.canonicalName
local name = self._rawData[1] or self._rawData.canonicalName
-- If the name already has "code" or "semaphore" in it, don't add it.
-- If the name already has "code" or "semaphore" in it, don't add it.
Line 71: Line 98:
return name
return name
end
end


function Script:makeCategoryLink()
function Script:makeCategoryLink()
Line 77: Line 103:
end
end


 
--[==[Returns the {{code|lua|wikipedia_article}} item in the language's data file, or else calls {{code|lua|Script:getCategoryName()}}.]==]
function Script:getWikipediaArticle()
function Script:getWikipediaArticle()
return self._rawData.wikipedia_article or self:getCategoryName()
return self._rawData.wikipedia_article or self:getCategoryName()
end
end


 
--[==[Returns the regex defining the script's characters from the language's data file.
This can be used to search for words consisting only of this script, but see the warning above.]==]
function Script:getCharacters()
function Script:getCharacters()
if self._rawData.characters then
if self._rawData.characters then
Line 91: Line 118:
end
end


 
--[==[Returns the number of characters in the text that are part of this script.
'''Note:''' You should never rely on text consisting entirely of the same script. Strings may contain spaces, punctuation and even wiki markup or HTML tags. HTML tags will skew the counts, as they contain Latin-script characters. So it's best to avoid them.]==]
function Script:countCharacters(text)
function Script:countCharacters(text)
if not self._rawData.characters then
if not self._rawData.characters then
return 0
return 0
-- Due to the number of Chinese characters, a different determination method is used when differentiating between traditional ("Hant") and simplified ("Hans") Chinese.
elseif self:getCode() == "Hant" or self:getCode() == "Hans" then
local charData, num = self:getCode() == "Hant" and require("Module:zh/data/ts/serialized") or require("Module:zh/data/st/serialized"), 0
charData = charData:sub(1, charData:len() / 2)
for char in text:gmatch("[\194-\244][\128-\191]*") do
if charData:find(char) then
num = num + 1
end
end
return num
else
else
local _, num = mw.ustring.gsub(text, "[" .. self._rawData.characters .. "]", "")
local _, num = mw.ustring.gsub(text, "[" .. self._rawData.characters .. "]", "")
Line 101: Line 139:
end
end


function Script:hasCapitalization()
return not not self._rawData.capitalized
end
function Script:hasSpaces()
return self._rawData.spaces ~= false
end
function Script:isTransliterated()
return self._rawData.translit ~= false
end
--[==[Returns true if the script is (sometimes) sorted by scraping page content, meaning that it is sensitive to changes in capitalization during sorting.]==]
function Script:sortByScraping()
return not not self._rawData.sort_by_scraping
end
--[==[Returns the text direction, if any. Currently, left-to-right scripts are unmarked, while most right-to-left scripts have direction specified as {{code|lua|"rtl"}} and Mongolian as {{code|lua|"down"}}.]==]
function Script:getDirection()
function Script:getDirection()
local direction = self._rawData.direction
return self._rawData.direction
if not direction then
return nil
else
return direction
end
end
end


Line 115: Line 166:
end
end


--[==[Returns {{code|lua|true}} if the script contains characters that require fixes to Unicode normalization under certain circumstances, {{code|lua|false}} if it doesn't.]==]
function Script:hasNormalizationFixes()
return not not self._rawData.normalizationFixes
end
--[==[Corrects discouraged sequences of Unicode characters to the encouraged equivalents.]==]
function Script:fixDiscouragedSequences(text)
if self:hasNormalizationFixes() and self._rawData.normalizationFixes.from then
local gsub = require("Module:string utilities").gsub
for i, from in ipairs(self._rawData.normalizationFixes.from) do
text = gsub(text, from, self._rawData.normalizationFixes.to[i] or "")
end
end
return text
end
-- Implements a modified form of Unicode normalization for instances where there are identified deficiencies in the default Unicode combining classes.
local function fixNormalization(text, self)
if self:hasNormalizationFixes() and self._rawData.normalizationFixes.combiningClasses then
local combiningClassFixes = self._rawData.normalizationFixes.combiningClasses
local charsToFix = table.concat(require("Module:table").keysToList(combiningClassFixes))
if require("Module:string utilities").match(text, "[" .. charsToFix .. "]") then
local codepoint, u = mw.ustring.codepoint, mw.ustring.char
-- Obtain the list of default combining classes.
local combiningClasses = mw.loadData("Module:scripts/data/combiningClasses")
-- For each character that needs fixing, find all characters with combining classes equal to or lower than its default class, but greater than its new class (i.e. intermediary characters).
for charToFix, newCombiningClass in pairs(combiningClassFixes) do
local intermediaryChars = {}
for character, combiningClass in pairs(combiningClasses) do
if newCombiningClass < combiningClass and combiningClass <= combiningClasses[codepoint(charToFix)] then
table.insert(intermediaryChars, u(character))
end
end
-- Swap the character with any intermediary characters that are immediately before it.
text = require("Module:string utilities").gsub(text, "([" .. table.concat(intermediaryChars) .. "]+)(" .. charToFix .. ")", "%2%1")
end
end
end
return text
end
function Script:toFixedNFC(text)
return fixNormalization(mw.ustring.toNFC(text), self)
end
function Script:toFixedNFD(text)
return fixNormalization(mw.ustring.toNFD(text), self)
end
function Script:toFixedNFKC(text)
return fixNormalization(mw.ustring.toNFKC(text), self)
end
function Script:toFixedNFKD(text)
return fixNormalization(mw.ustring.toNFKD(text), self)
end


function Script:toJSON()
function Script:toJSON()
if not self._type then
self:hasType()
end
local types = {}
for type in pairs(self._type) do
table.insert(types, type)
end
local ret = {
local ret = {
canonicalName = self:getCanonicalName(),
canonicalName = self:getCanonicalName(),
categoryName = self:getCategoryName("nocap"),
categoryName = self:getCategoryName("nocap"),
code = self._code,
code = self:getCode(),
otherNames = self:getOtherNames(true),
otherNames = self:getOtherNames(true),
aliases = self:getAliases(),
aliases = self:getAliases(),
varieties = self:getVarieties(),
varieties = self:getVarieties(),
type = self:getType(),
type = types,
direction = self:getDirection(),
direction = self:getDirection(),
characters = self:getCharacters(),
characters = self:getCharacters(),
parent = self:getParent(),
parent = self:getParent(),
systems = self._rawData.systems or {},
systems = self:getSystemCodes(),
wikipediaArticle = self._rawData.wikipedia_article,
wikipediaArticle = self._rawData.wikipedia_article,
}
}
Line 134: Line 249:
return require("Module:JSON").toJSON(ret)
return require("Module:JSON").toJSON(ret)
end
end


Script.__index = Script
Script.__index = Script
 
 
function export.makeObject(code, data, useRequire)
function export.makeObject(code, data)
return data and setmetatable({_rawData = data, _code = code}, Script) or nil
return data and setmetatable({ _rawData = data, _code = code }, Script) or nil
end
end


-- Temporary aliases from canonicalized names to (existing) anomalous names. Once we have converted everything we will
-- rename the scripts and remove the alias code.
local scriptAliases = {
-- scripts already renamed; we now alias the old names to the new ones
["IPAchar"] = "Ipach",
["musical"] = "Music",
["Ruminumerals"] = "Rumin",
["polytonic"] = "Polyt",
["Latinx"] = "Latnx",
}


function export.getByCode(code, paramForError, disallowNil)
--[==[Finds the script whose code matches the one provided. If it exists, it returns a {{code|lua|Script}} object representing the script. Otherwise, it returns {{code|lua|nil}}, unless <span class="n">paramForError</span> is given, in which case an error is generated. If <code class="n">paramForError</code> is {{code|lua|true}}, a generic error message mentioning the bad code is generated; otherwise <code class="n">paramForError</code> should be a string or number specifying the parameter that the code came from, and this parameter will be mentioned in the error message along with the bad code.]==]
function export.getByCode(code, paramForError, disallowNil, useRequire)
if code == nil and not disallowNil then
if code == nil and not disallowNil then
return nil
return nil
end
end
local retval = export.makeObject(code, mw.loadData("Module:scripts/data")[code])
code = scriptAliases[code] or code
if not retval and paramForError then
require("Module:languages").err(code, paramForError, "script code", nil, "not real lang")
end
return retval
end
 
function export.getByCanonicalName(name)
local code = mw.loadData("Module:scripts/by name")[name]
if not code then
local data
return nil
if useRequire then
data = require("Module:scripts/data")[code]
else
data = mw.loadData("Module:scripts/data")[code]
end
end
return export.makeObject(code, mw.loadData("Module:scripts/data")[code])
local retval = export.makeObject(code, data, useRequire)
end
 
-- Find the best script to use, based on the characters of a string.
-- If forceDetect is set, run the detection algorithm even if there's only one
-- possible script; in that case, if the text isn't in the script, the return
-- value will be None.
function export.findBestScript(text, lang, forceDetect)
if not text or not lang or not lang.getScripts then
return export.getByCode("None")
end
local scripts = lang:getScripts()
if not retval and paramForError then
require("Module:languages/error")(code, paramForError, "script code", nil, "not real lang")
if not scripts[2] and not forceDetect then
return scripts[1]
end
end
--[=[
return retval
Remove any HTML entities; catfix function in [[Module:utilities]]
adds tagging to a no-break space (&nbsp;), which contains Latin characters;
hence Latin was returned as the script if "Latn" is one of the language's scripts.
]=]
text = string.gsub(text, "&[a-zA-Z0-9]+;", "")
-- Try to match every script against the text,
-- and return the one with the most matching characters.
local bestcount = 0
local bestscript = nil
-- Get length of text minus any spacing or punctuation characters.
-- Counting instances of UTF-8 character pattern is faster than mw.ustring.len.
local _, length = string.gsub(mw.ustring.gsub(text, "[%s%p]+", ""), "[\1-\127\194-\244][\128-\191]*", "")
if length == 0 then
return export.getByCode("None")
end
for i, script in ipairs(scripts) do
local count = script:countCharacters(text)
if count >= length then
return script
end
if count > bestcount then
bestcount = count
bestscript = script
end
end
if bestscript then
return bestscript
end
-- No matching script was found. Return "None".
return export.getByCode("None")
end
end


-- Copied from [[Module:Unicode data]].
function export.getByCanonicalName(name, useRequire)
local floor = math.floor
local code
local function binaryRangeSearch(codepoint, ranges)
if useRequire then
local low, mid, high
code = require("Module:scripts/by name")[name]
low, high = 1, ranges.length or require "Module:table".length(ranges)
else
while low <= high do
code = mw.loadData("Module:scripts/by name")[name]
mid = floor((low + high) / 2)
local range = ranges[mid]
if codepoint < range[1] then
high = mid - 1
elseif codepoint <= range[2] then
return range, mid
else
low = mid + 1
end
end
return nil, mid
end
 
-- Copied from [[Module:Unicode data]].
local function linearRangeSearch(codepoint, ranges)
for i, range in ipairs(ranges) do
if codepoint < range[1] then
break
elseif codepoint <= range[2] then
return range
end
end
end
return export.getByCode(code, nil, nil, useRequire)
end
end


local function compareRanges(range1, range2)
--[==[
return range1[1] < range2[1]
end
 
-- Save previously used codepoint ranges in case another character is in the
-- same range.
local rangesCache = {}
 
--[=[
Takes a codepoint or a character and finds the script code (if any) that is
Takes a codepoint or a character and finds the script code (if any) that is
appropriate for it based on the codepoint, using the data module
appropriate for it based on the codepoint, using the data module
Line 268: Line 310:
is in the list of individual characters, or if it is in one of the defined
is in the list of individual characters, or if it is in one of the defined
ranges in the 4096-character block that it belongs to, else returns "None".
ranges in the 4096-character block that it belongs to, else returns "None".
]=]
]==]
local charToScriptData
function export.charToScript(char)
function export.charToScript(char)
charToScriptData = charToScriptData or mw.loadData("Module:scripts/recognition data")
return require("Module:scripts/charToScript").charToScript(char)
local t = type(char)
local codepoint
if t == "string" then
local etc
codepoint, etc = mw.ustring.codepoint(char, 1, 2)
if etc then
error("bad argument #1 to 'charToScript' (expected a single character)")
end
elseif t == "number" then
codepoint = char
else
error(("bad argument #1 to 'charToScript' (expected string or a number, got %s)")
:format(t))
end
 
local individualMatch = charToScriptData.individual[codepoint]
if individualMatch then
return individualMatch
else
local range
if rangesCache[1] then
range = linearRangeSearch(codepoint, rangesCache)
if range then
return range[3]
end
end
 
local index = floor(codepoint / 0x1000)
 
range = linearRangeSearch(index, charToScriptData.blocks)
if not range and charToScriptData[index] then
range = binaryRangeSearch(codepoint, charToScriptData[index])
if range then
table.insert(rangesCache, range)
table.sort(rangesCache, compareRanges)
end
end
return range and range[3] or "None"
end
end
end


--[==[Returns the code for the script that has the greatest number of characters in <code>text</code>. Useful for script tagging text that is unspecified for language. Uses [[Module:scripts/recognition data]] to determine a script code for a character language-agnostically.]==]
function export.findBestScriptWithoutLang(text)
function export.findBestScriptWithoutLang(text)
local scripts = {}
return require("Module:scripts/charToScript").findBestScriptWithoutLang(text)
for character in text:gmatch("[%z\1-\127\194-\244][\128-\191]*") do
local script = export.charToScript(character)
scripts[script] = (scripts[script] or 0) + 1
end
local bestScript
local greatestCount = 0
for script, count in pairs(scripts) do
if count > greatestCount then
bestScript = script
greatestCount = count
end
end
return bestScript
end
end


return export
return export

Revision as of 14:01, 29 July 2023



local export = {}

local Script = {}

--[==[Returns the script code of the language. Example: {{code|lua|"Cyrl"}} for Cyrillic.]==]
function Script:getCode()
	return self._code
end

--[==[Returns the canonical name of the script. This is the name used to represent that script on Wiktionary. Example: {{code|lua|"Cyrillic"}} for Cyrillic.]==]
function Script:getCanonicalName()
	return self._rawData[1] or self._rawData.canonicalName
end

--[==[Returns the display form of the script. For scripts, this is the same as the value returned by <code>:getCategoryName("nocap")</code>, i.e. it reads "NAME script" (e.g. {{code|lua|"Arabic script"}}). For regular and etymology languages, this is the same as the canonical name, and for families, it reads "NAME languages" (e.g. {{code|lua|"Indo-Iranian languages"}}). The displayed text used in <code>:makeCategoryLink</code> is always the same as the display form.]==]
function Script:getDisplayForm()
	return self:getCategoryName("nocap")
end

function Script:getOtherNames(onlyOtherNames)
	return require("Module:language-like").getOtherNames(self, onlyOtherNames)
end

function Script:getAliases()
	return self._rawData.aliases or {}
end

function Script:getVarieties(flatten)
	return require("Module:language-like").getVarieties(self, flatten)
end

--[==[Returns the parent of the script. Example: {{code|lua|"Latn"}} for {{code|lua|"Latnx"}} and {{code|lua|"Arab"}} for {{code|lua|"fa-Arab"}}. It returns {{code|lua|"top"}} for scripts without a parent, like {{code|lua|"Latn"}}, {{code|lua|"Grek"}}, etc.]==]
function Script:getParent()
	return self._rawData.parent
end

function Script:getSystemCodes()
	if not self._systemCodes then
		if type(self._rawData[2]) == "table" then
			self._systemCodes = self._rawData[2]
		elseif type(self._rawData[2]) == "string" then
			self._systemCodes = mw.text.split(self._rawData[2], "%s*,%s*")
		else
			self._systemCodes = {}
		end
	end
	return self._systemCodes
end

function Script:getSystems()
	if not self._systemObjects then
		local m_systems = require("Module:writing systems")
		self._systemObjects = {}
		
		for _, ws in ipairs(self:getSystemCodes()) do
			table.insert(self._systemObjects, m_systems.getByCode(ws))
		end
	end
	
	return self._systemObjects
end

--function Script:getAllNames()
--	return self._rawData.names
--end

--[==[Given a list of types as strings, returns true if the script has all of them. Possible types are explained in [[Module:scripts/data]].]==]
function Script:hasType(...)
	if not self._type then
		self._type = {script = true}
		if self._rawData.type then
			for _, type in ipairs(mw.text.split(self._rawData.type, "%s*,%s*")) do
				self._type[type] = true
			end
		end
	end
	for _, type in ipairs{...} do
		if not self._type[type] then
			return false
		end
	end
	return true
end

--[==[Returns the name of the main category of that script. Example: {{code|lua|"Cyrillic script"}} for Cyrillic, whose category is at [[:Category:Cyrillic script]].
Unless optional argument <code>nocap</code> is given, the script name at the beginning of the returned value will be capitalized. This capitalization is correct for category names, but not if the script name is lowercase and the returned value of this function is used in the middle of a sentence. (For example, the script with the code <code>Semap</code> has the name <code>"flag semaphore"</code>, which should remain lowercase when used as part of the category name [[:Category:Translingual letters in flag semaphore]] but should be capitalized in [[:Category:Flag semaphore templates]].) If you are considering using <code>getCategoryName("nocap")</code>, use <code>getDisplayForm()</code> instead.]==]
function Script:getCategoryName(nocap)
	local name = self._rawData[1] or self._rawData.canonicalName
	
	-- If the name already has "code" or "semaphore" in it, don't add it.
	-- No names contain "script".
	if not name:find("[Cc]ode$") and not name:find("[Ss]emaphore$") then
		name = name .. " script"
	end
	if not nocap then
		name = mw.getContentLanguage():ucfirst(name)
	end
	return name
end

function Script:makeCategoryLink()
	return "[[:Category:" .. self:getCategoryName() .. "|" .. self:getDisplayForm() .. "]]"
end

--[==[Returns the {{code|lua|wikipedia_article}} item in the language's data file, or else calls {{code|lua|Script:getCategoryName()}}.]==]
function Script:getWikipediaArticle()
	return self._rawData.wikipedia_article or self:getCategoryName()
end

--[==[Returns the regex defining the script's characters from the language's data file.
This can be used to search for words consisting only of this script, but see the warning above.]==]
function Script:getCharacters()
	if self._rawData.characters then
		return self._rawData.characters
	else
		return nil
	end
end

--[==[Returns the number of characters in the text that are part of this script.
'''Note:''' You should never rely on text consisting entirely of the same script. Strings may contain spaces, punctuation and even wiki markup or HTML tags. HTML tags will skew the counts, as they contain Latin-script characters. So it's best to avoid them.]==]
function Script:countCharacters(text)
	if not self._rawData.characters then
		return 0
	-- Due to the number of Chinese characters, a different determination method is used when differentiating between traditional ("Hant") and simplified ("Hans") Chinese.
	elseif self:getCode() == "Hant" or self:getCode() == "Hans" then
		local charData, num = self:getCode() == "Hant" and require("Module:zh/data/ts/serialized") or require("Module:zh/data/st/serialized"), 0
		charData = charData:sub(1, charData:len() / 2)
		for char in text:gmatch("[\194-\244][\128-\191]*") do
			if charData:find(char) then
				num = num + 1
			end
		end
		return num
	else
		local _, num = mw.ustring.gsub(text, "[" .. self._rawData.characters .. "]", "")
		return num
	end
end

function Script:hasCapitalization()
	return not not self._rawData.capitalized
end

function Script:hasSpaces()
	return self._rawData.spaces ~= false
end

function Script:isTransliterated()
	return self._rawData.translit ~= false
end

--[==[Returns true if the script is (sometimes) sorted by scraping page content, meaning that it is sensitive to changes in capitalization during sorting.]==]
function Script:sortByScraping()
	return not not self._rawData.sort_by_scraping
end

--[==[Returns the text direction, if any. Currently, left-to-right scripts are unmarked, while most right-to-left scripts have direction specified as {{code|lua|"rtl"}} and Mongolian as {{code|lua|"down"}}.]==]
function Script:getDirection()
	return self._rawData.direction
end


function Script:getRawData()
	return self._rawData
end

--[==[Returns {{code|lua|true}} if the script contains characters that require fixes to Unicode normalization under certain circumstances, {{code|lua|false}} if it doesn't.]==]
function Script:hasNormalizationFixes()
	return not not self._rawData.normalizationFixes
end

--[==[Corrects discouraged sequences of Unicode characters to the encouraged equivalents.]==]
function Script:fixDiscouragedSequences(text)
	if self:hasNormalizationFixes() and self._rawData.normalizationFixes.from then
		local gsub = require("Module:string utilities").gsub
		for i, from in ipairs(self._rawData.normalizationFixes.from) do
			text = gsub(text, from, self._rawData.normalizationFixes.to[i] or "")
		end
	end
	return text
end

-- Implements a modified form of Unicode normalization for instances where there are identified deficiencies in the default Unicode combining classes.
local function fixNormalization(text, self)
	if self:hasNormalizationFixes() and self._rawData.normalizationFixes.combiningClasses then
		local combiningClassFixes = self._rawData.normalizationFixes.combiningClasses
		local charsToFix = table.concat(require("Module:table").keysToList(combiningClassFixes))
		if require("Module:string utilities").match(text, "[" .. charsToFix .. "]") then
			local codepoint, u = mw.ustring.codepoint, mw.ustring.char
			-- Obtain the list of default combining classes.
			local combiningClasses = mw.loadData("Module:scripts/data/combiningClasses")
			-- For each character that needs fixing, find all characters with combining classes equal to or lower than its default class, but greater than its new class (i.e. intermediary characters).
			for charToFix, newCombiningClass in pairs(combiningClassFixes) do
				local intermediaryChars = {}
				for character, combiningClass in pairs(combiningClasses) do
					if newCombiningClass < combiningClass and combiningClass <= combiningClasses[codepoint(charToFix)] then
						table.insert(intermediaryChars, u(character))
					end
				end
				-- Swap the character with any intermediary characters that are immediately before it.
				text = require("Module:string utilities").gsub(text, "([" .. table.concat(intermediaryChars) .. "]+)(" .. charToFix .. ")", "%2%1")
			end
		end
	end
	return text
end

function Script:toFixedNFC(text)
	return fixNormalization(mw.ustring.toNFC(text), self)
end

function Script:toFixedNFD(text)
	return fixNormalization(mw.ustring.toNFD(text), self)
end

function Script:toFixedNFKC(text)
	return fixNormalization(mw.ustring.toNFKC(text), self)
end

function Script:toFixedNFKD(text)
	return fixNormalization(mw.ustring.toNFKD(text), self)
end

function Script:toJSON()
	if not self._type then
		self:hasType()
	end
	local types = {}
	for type in pairs(self._type) do
		table.insert(types, type)
	end
	
	local ret = {
		canonicalName = self:getCanonicalName(),
		categoryName = self:getCategoryName("nocap"),
		code = self:getCode(),
		otherNames = self:getOtherNames(true),
		aliases = self:getAliases(),
		varieties = self:getVarieties(),
		type = types,
		direction = self:getDirection(),
		characters = self:getCharacters(),
		parent = self:getParent(),
		systems = self:getSystemCodes(),
		wikipediaArticle = self._rawData.wikipedia_article,
	}
	
	return require("Module:JSON").toJSON(ret)
end

Script.__index = Script
	
function export.makeObject(code, data, useRequire)
	return data and setmetatable({_rawData = data, _code = code}, Script) or nil
end

-- Temporary aliases from canonicalized names to (existing) anomalous names. Once we have converted everything we will
-- rename the scripts and remove the alias code.
local scriptAliases = {
	-- scripts already renamed; we now alias the old names to the new ones
	["IPAchar"] = "Ipach",
	["musical"] = "Music",
	["Ruminumerals"] = "Rumin",
	["polytonic"] = "Polyt",
	["Latinx"] = "Latnx",
}

--[==[Finds the script whose code matches the one provided. If it exists, it returns a {{code|lua|Script}} object representing the script. Otherwise, it returns {{code|lua|nil}}, unless <span class="n">paramForError</span> is given, in which case an error is generated. If <code class="n">paramForError</code> is {{code|lua|true}}, a generic error message mentioning the bad code is generated; otherwise <code class="n">paramForError</code> should be a string or number specifying the parameter that the code came from, and this parameter will be mentioned in the error message along with the bad code.]==]
function export.getByCode(code, paramForError, disallowNil, useRequire)
	if code == nil and not disallowNil then
		return nil
	end
	code = scriptAliases[code] or code
	
	local data
	if useRequire then
		data = require("Module:scripts/data")[code]
	else
		data = mw.loadData("Module:scripts/data")[code]
	end
	
	local retval = export.makeObject(code, data, useRequire)
	
	if not retval and paramForError then
		require("Module:languages/error")(code, paramForError, "script code", nil, "not real lang")
	end
	
	return retval
end

function export.getByCanonicalName(name, useRequire)
	local code
	if useRequire then
		code = require("Module:scripts/by name")[name]
	else
		code = mw.loadData("Module:scripts/by name")[name]
	end
	
	return export.getByCode(code, nil, nil, useRequire)
end

--[==[
	Takes a codepoint or a character and finds the script code (if any) that is
	appropriate for it based on the codepoint, using the data module
	[[Module:scripts/recognition data]]. The data module was generated from the
	patterns in [[Module:scripts/data]] using [[Module:User:Erutuon/script recognition]].

	Converts the character to a codepoint. Returns a script code if the codepoint
	is in the list of individual characters, or if it is in one of the defined
	ranges in the 4096-character block that it belongs to, else returns "None".
]==]
function export.charToScript(char)
	return require("Module:scripts/charToScript").charToScript(char)
end

--[==[Returns the code for the script that has the greatest number of characters in <code>text</code>. Useful for script tagging text that is unspecified for language. Uses [[Module:scripts/recognition data]] to determine a script code for a character language-agnostically.]==]
function export.findBestScriptWithoutLang(text)
	return require("Module:scripts/charToScript").findBestScriptWithoutLang(text)
end

return export